The translation of scientific texts from Greek and Arabic in the twelfth and thirteenth centuries is both a symptom and a cause of one of the greatest shifts in Western science, comparable in importance with the parallel importation of scientific works into Arabic in Baghdad in the ninth century and the Scientific Revolution in seventeenth-century Europe.

The motivation for the translations was the perceived lacunae in Latin scientific education. There were three principal areas in which the Latins were felt to be particularly lacking. The first was mathematics, especially geometry and astronomy. These were two of the seven liberal arts, which had formed the framework for the rhetorically-based Latin education since Late Antiquity and had been revived first by Gerbert d’Aurillac at the turn of the millennium, then, more generally, in the twelfth century. The other five liberal arts were well represented, mostly thanks to the translations of Boethius (d. 529), who had sought to provide Latin readers with a complete curriculum of Greek studies. But geometry and astronomy lacked comprehensive text books.

For geometry Boethius had only translated a small portion of Euclid’s Elements. Stephen the Philosopher, working in Antioch in the early twelfth century, lamented the poor knowledge of geometry among the Latins, and John of Salisbury reckoned that the only place where the study flourished was in (Islamic) Spain. Thus, the pioneer in the twelfth-century translating movement, Adelard of Bath, devoted a text to the description of the seven liberal arts (his On the Same and the Different), but also made the first translation of the Elements from Arabic. For astronomy no translation by Boethius has survived, but the Latins were aware that the most important Greek text was the Almagest of Ptolemy, and it is specifically for this text that Gerard of Cremona is said to have gone to Toledo, while the slightly earlier translator of the Almagest from Greek had rushed to Sicily from Salerno because he heard that a copy of the Almagest had just been brought there from Constantinople. How the new translations filled gaps in the traditional curriculum can be seen in the important ‘Library of the Seven Liberal Arts’ put together by Thierry of Chartres (d. 1151) in the early 1140s (the Heptateuchon) which included both Euclid’s Elements translated from Arabic and astronomical tables by the ninth-century Arabic mathematician al-Khwarizmi.

The second lacuna was in physics (i.e., the investigation of the workings of nature). In this case we are dealing with a subject that was not one of the rhetorically-based seven liberal arts, but had been part of the curriculum in philosophy in ancient Alexandria, and which continued to be taught (with interruptions and revivals) in Byzantium and (more continuously) in the Islamic world. In the West the interest in physics pre-dates the knowledge of the texts of this curriculum, and is based on “natural questions” and Late Antique Latin works such as Calcidius’ translation of Plato’s Timaeus and Macrobius’s commentary on the Somnium Scipionis of Cicero. However, during the course of the twelfth century, the fact that Aristotle, aside from being an authority on logic, was also the authority on natural science became known. Burgundio of Pisa, James of Venice, and William of Moerbeke were directly aware of the corpus of Aristotle’s natural science from contemporary Greek scholars who were teaching and writing commentaries on it, and Gerard of Cremona knew the same texts in Arabic and as described by Alfarabi in the Catalogue of Sciences, which he translated.

The third area was medicine. In this subject Galen was the master, and the titles of the sixteen basic texts of his that formed the curriculum at Alexandria were known. Constantine the African (d. before 1098/9) listed them in his preface to the Pantegni and translated two of them from Arabic; several more were translated by Burgundio of Pisa and Gerard of Cremona. If the goals of the translators were to restore the ancient learning of Euclid, Ptolemy, Aristotle, and Galen, they had two sources: the centres of Greek and of Arabic learning. The Byzantine Greeks on the whole had preserved the ancient texts without substantially altering them; there had been little scientific development, and a twelfth-century commentary on a work by Aristotle could easily be taken for a second-century commentary on the same work. Among the Greeks, therefore, the Latins sought and could find their copies of Ptolemy’s Almagest, Aristotle’s libri naturales, and Galen’s works, just as Arabic scholars had done in the ninth century. For the Latins, manuscripts that the Greeks could provide were more important than scholarly expertise. For the most part, the interpretations of the ancient texts were ancient themselves: those of Proclus and Themistius, Alexander of Aphrodisias, and John Philoponus.

Among the Arabs, however, the Latins could and did find these same Greek works, but they were also confronted with the results of a tradition of scholarship which had not only absorbed new elements from other cultures (particularly those of India and Persia), but had also developed, refined, and changed the learning of the ancients. Thus, the astronomical models of Ptolemy had to contend with those of Indian astronomers, and his measurements of the movements of the planets were repeatedly corrected and often replaced, ever since the official “testing” sponsored by the caliph al-Mamun in the early ninth century. Aristotle’s libri naturales together with the ancient commentaries on them were transmitted, but Alfarabi, Avempace (Ibn Bajja, late eleventh century to ca. 1139), and Averroes wrote new commentaries, and Avicenna recast the whole of Aristotle’s philosophy. Perhaps even more radical was the replacement of the original works of Galen by new texts on medicine, each generation of Arabic doctors trying to improve on the work of their predecessors.
Arabic learning, then, differed from the Greek in that it resided in masters as much as in books. Adelard and Stephen the Philosopher both refer to their Arabic magistri, which they encountered in the Principality of Antioch.

Other translators benefited from the diaspora of Jewish scholars who had cultivated Arabic learning, following the expulsion of the Jews from Islamic Spain by the Almohads in 1160 (in much the same way as Renaissance Italian scholars benefited from the exile of Jewish scholars from Spain in 1491). In some cases translators appeared to have used those scientific works promoted by their Arabic masters: an example is Constantine the African, who transmitted the medical tradition of his masters in Qairawan, and Gundissalinus, who translated works of Avicenna, Algazel, and Avicebron, the favored authors of his collaborator, Avendauth (Abraham Ibn Daud). Gerard of Cremona, while translating several important works by Arabic scholars (especially in the field of medicine), appears to have made a more deliberate effort to recover the ancient texts from amongst the Arabs. But his translating activity coincides with a reaction against “modern” developments also on the part of a group of Islamic scholars in Córdoba (the chief of them being Ibn Rushd), who tried to restore a pure Aristotle both for natural science and for astronomy.

For translations to be made, either the Latin scholar must go in search of the texts, or the texts must be sent or brought to the Latin centre of learning. Scholars might be part of the booty captured by pirates, as stated in one story concerning how Constantine the African arrived in Salerno (MS British Library, Sloane 2426, fols 8r–v). Other scholars arrived in Latin centres as the result of persecution or religious differences, such as the Arabic-speaking Christians (Mozarabs) who left Islamic Toledo for the Christian North of Spain in considerable numbers in the ninth century. After its reconquest, Toledo was the natural place of refuge for the Jews and Mozarabs who were driven out of Islamic Spain in the mid-twelfth century by the Almohad regime, and the flourishing study of the works of Averroes outside Islamic territory in the early thirteenth century may have been partly caused by intolerance towards the philosophical sciences among the later Almohad jurists.

But while enforced exile and the closing of doors brought about some contacts, the opening up of the Mediterranean to the Latins through conquest and trade brought more. It is no coincidence that the translation movement took off soon after the reconquest of Toledo, which opened up the heart of Islamic Spain (1085), the Norman conquest of Sicily with its Greek and Arabic-speaking population (1072–91), and the fall of Antioch, which opened up the Islamic and Greek culture of the Eastern Mediterranean (1098). The attempts at reunifying the Greek and Latin Churches also brought East and West closer together, and resulted in scientific translations as well as theological writings. Political leaders exchanged scholars as well as books: the Ayyubid Sultan, al-Malik al-Salih (1232-9) sent Frederick II one of the most distinguished Muslim scholars, Siraj al-Din al-Urmawi (d. 1283), to help him interpret Arabic logic.

More important over a longer period were the quarters set up by the Pisans and Venetians in cities throughout the Mediterranean, which, aside from their commercial function, offered opportunities for Latin scholars to work in the midst of Arabs and Greeks. Pisan quarters were set up in Antioch in 1108 and Constantinople in 1111. While James of Venice is attested in the Pisan quarter of Constantinople in 1136, Leonard of Pisa (Fibonacci) acquired his knowledge of Arabic mathematics in the Pisan debot of Bougie (Algeria).

In the case of the mathematical translations made from Arabic into Latin a significant role may have been played by a single royal library, that of the Banu Hud in Saragossa. Yusuf al-Mutaman ibn Hud, king from 1081 to 1085, had written a comprehensive book on geometry, al-Istikmal, which exploited a large collection of Greek and Arabic works on the subject. In 1110 the Banu Hud were driven out of Saragossa by the Almoravids, and settled in the fortress of Rueda de Jalón in Aragon. But they took their library with them, for Hugo of Santalla specifically says (in a rare example of the mention of a source) that his patron, Michael bishop of Tarazona, acquired the Arabic manuscript of a work on astronomical tables in this library (“in Rotensi armario”).

In 1140, the last of the Banu Hud, Abu Jafar Amad III ayf al-Dawla (d. 1146), exchanged his property in Rueda de Jalón for a house in the cathedral quarter in Toledo. It may be no accident that only after this move are mathematical works by Greek and Arabic authorities translated from Arabic in Toledo. Ever since the publication of a fundamental article by Valentin Rose in 1874, it has been thought that the archbishops of Toledo were directly responsible for sponsoring and promoting a “school of translators” as a quasi-university. The evidence is not decisive, however, since only two translations are dedicated to archbishops. Nevertheless, cathedral clergy appear to have played the leading role in the Toledan translation movement, and the locus of this activity must have been in the Cathedral quarter and the adjacent Frankish quarter of the city, the only districts in which foreigners and Latin learning were dominant—also, apparently, where the remains of the library of the Banu Hud were located, as we have seen.

The process by which this translation occurred can, however, take a variety of forms.
In some cases it involved the fixing in Latin of a text transmitted orally. That whole books could be dictated by scholars (especially those of Jewish origin) is not implausible. However, most medieval translations imply the existence of a written text in the original language, even if this text has been interpreted orally for a Latin writer. The source text need not be written in a conventional book form. It may be a schematic diagram of a sheep’s shoulder blade with the significance of each of the parts written on it, or an astrolabe with the names of prominent fixed stars and various figures inscribed on it. Hindu-Arabic numerals at first seem to have spread through Europe on the beads of the “Gerbertian abacus.” Nor is the text in the book merely the author’s words: it may include diagrams and illustrations, as well as marginalia, and its layout may influence the choices of the translator. It matters, then, whether the Latin translator has confronted a text in the original manuscript directly, or has merely relied on an oral interpretation of that text. Both situations occurred.
Oral interpretation is best attested in respect to Arabic texts. A well-known description of the process is given by Avendauth, in the dedication of his translation of Avicenna’s De anima to the archbishop of Toledo. Here you have the book translated from Arabic. I took the lead and translated the words one at a time into the vernacular language, and Archdeacon Domenicus turned them one at a time into Latin. Since this translation was made in Toledo, the “vernacular language” could have been either the Arabic dialect spoken in Toledo, or a Romance language. According to Roger Bacon, many people understood the spoken forms of Greek, Arabic, and Hebrew, even if they could not read and write the languages. Other translators are also said to have had the help of Arabic speakers, whether Mozarabs (such as the “Galippus” who helped Gerard of Cremona), or Jews (such as the “Salomon Avenraza” whom Alfred of Shareshill mentions as his teacher, and the “Abuteus levita” who helped Michael Scot).

Evidence of the use of dictionaries and grammars is less easy to find. The earliest Arabic grammar composed in a Western language is the one for the Arabic dialect of Granada written in 1505 by Pedro de Alcalà. Its purpose was to aid the conversion of the Arabic speaking population, and much the same aim may underlie the two extensive Arabic-Latin/Latin-Arabic glossaries of the Middle Ages: that so-called “Leiden glossary” and the Vocabulista in Arabico edited by Celestino Schiaparelli (Florence, 1871). Potentially more helpful are the glossaries of technical terms attached to certain texts of astronomy, astrology and medicine. The most extensive of these is the Synonyma of Simon of Genoa, a doctor at the papal court at the very end of the thirteenth century, who explained in detail the Greek and Arabic words in the medical vocabulary of the Latins; he included some notes on Greek and Arabic phonology. These technical glossaries, however, were primarily for the use of readers, unfamiliar with terms which had simply been transliterated from Arabic or Greek.

Literal translation soon became the norm for scientific and philosophical works. It’s aim was to represent the words of the original author so accurately as to make the original texts redundant. Such a view is already apparent in the words of Boethius, that, “through the integrity of a completely full translation, no Greek literature is found to be needed any longer.” Boethius’s aim for a “transfer of learning” (translatio studii) from Greek into Latin in the sixth century was cut short by his untimely death and lack of immediate successors, but it was renewed and largely fulfilled in the twelfth to thirteenth centuries. The rationale of a translatio studii is that scientific culture passes from one people to another; once in possession of that culture, the receiving people has no need to return to its source.

But it is probably truer to say that we are dealing here not so much with a transfer of learning as with the internationalism of scientific learning. Already Adelard of Bath had compared the world with a body in which different parts have been assigned different functions: in the same way different parts of the world are fertile with different disciplines and what “the [world] soul is unable to effect in a single part of the world, she brings about within its totality.” In the next century, Theodore of Antioch went from Christian Antioch to Muslim Mosul where he studied the works of Alfarabi, Avicenna, Euclid, and Ptolemy with the foremost Islamic scholar, Kamal-al-Din ibn Yunus (1156–1242); subsequently he studied medicine in Baghdad, before serving a Seljuk ruler of Konya in Asia Minor, an Armenian regent and finally a Christian Emperor in Sicily. Another student of Kamal-al-Din, al-Urmawi, also spent time at the same Emperor’s court, writing a book of logic for him. A Jewish scholar, Juda b. Salomon ha-Cohen, corresponded in Arabic from Toledo with the Emperor Frederick II’s “philosopher” on questions concerning geometry, and Frederick himself sent questions on mathematics and philosophy to Arabic scholars throughout the Mediterranean.

Later in the same century, opinions of Thomas Aquinas were incorporated into an Arabic apologia for Christianity written by a Mozarab in Toledo and an anonymous scholar wrote a Greek introduction to Aristotle’s libri naturales in Sicily or South Italy on the basis of Averroistic texts being taught in Paris. Meanwhile, it has been suggested, astronomical information from the observatory of Maragha in the Mongol realms arrived at the Spanish court of Alfonso X. That such scholarly exchanges and intellectual traffic were possible testifies to the fact that, at least by the second quarter of the thirteenth century, the Jewish and Islamic world shared with Christendom a common knowledge of science and philosophy; a commonwealth of scholars had come into being, which transcended political and linguistic borders. That such a state could come about is due, in no small measure, to the achievement of translators who raised the scientific culture of each linguistic group to the point where all shared the same level of excellence.

This text has been read by the author during the conference Beyond Orientalism and Occidentalism, organized by Reset-Dialogues on Civilizations and held in Cairo, Egypt, March 4th/6th 2006.